Skip to main content
All Collections🤖 Flobot FeaturesAI OperationOCR
What You Can & Cannot Do When Extracting Text from Images & PDFs
What You Can & Cannot Do When Extracting Text from Images & PDFs

As of Aug 8, 2024

Yoom Customer Service avatar
Written by Yoom Customer Service
Updated this week

📝 Overview

In Yoom, you can use the AI actions to extract text from images or PDF files.

it is possible to extract text information from images or PDF files.

This guide explains what can and cannot be extracted based on different conditions.

☝️ Note:

This feature is only available in the Business & Enterprise Plan.


💡 Important Notes

As AI technology continues to improve, some tasks that are not currently possible may become feasible in the future. The details provided below may change over time.

The examples and extraction conditions mentioned are based on current testing, so results may vary if the format of the image or file changes. We recommend testing with your specific files or images to see how well the extraction works.


🙌 When Text Extraction is Possible

You can successfully extract text if these conditions are met:

  • Item names are clearly defined in a single image or file

    • Example: Address, phone number, name, amount, etc.

      List the specific item names you want to extract in the designated field.

  • The data in multi-page file is unique

    • Example: The names in the "Name" field are all different.

      If you're working with multi-page file, use keywords like "List of names" to help extract the data.

  • The File is Within the 30,000-character limit
    The action text extraction can only process files that are up to 30,000 characters or less.

  • Tabular Data with unique entries

    • Example: A table with unique values under each column.

      For these type of files, set keywords like "List of company names" to extract the data.

😰 When Extraction May Be Difficult

It can be harder to extract text in these situations:

  • Item names are not clearly defined

    Example: If the item names are missing or unclear, the AI might struggle to identify the information.

  • The data in multi-page file is not unique

    Example: If there are duplicate names in the "Name" field, even after specifying "List of names", some duplicates might be excluded, resulting in incomplete extraction.

  • The File is exceeds 30,000 characters
    Files longer than 30,000 characters can't be processed.

  • Broad extraction requests

    Using vague requests like "extract all text" may lead to incomplete or inaccurate extraction. It's best to be specific about the text you want to extract.

  • Vertical text

    If text is written vertically, it may not be extracted properly, especially if it covers a large area.

  • Figures or Non-Text information

    The AI can extract text within images or figures, but it cannot extract the figures themselves.

  • A data file in tabular format with excessive gaps

    If the data file contains blanks or if there are gaps of two or more lines within the table, the extraction may not work as expected.

That’s an overview of what can and cannot be done when extracting text from images and PDFs in Yoom. 🎉

Search Keywords

Image, PDF, reading, OCR, important notes, AI, limitations, text extraction

Did this answer your question?